Signal Degradation Monitoring

NVIDIA MLNX-OS User Manual v3.10.5002

A system can monitor the Bit Error Rate (BER) in order to ensure a quality of the link. As long as BER observed by the MACLRH layer is low enough, the rate of packet loss is low enough to allow successful operation of the applications running on top of the network.

The system continuously monitors the link BER and compares it to BER limits, when limits are crossed the system can generate an event indicating that link quality is degraded to the network operator that can take preemptive actions or even disable the low quality link. The BER configuration threshold for No-FEC and Post-FEC configured for 10^-10 will cause the trap to occur.

When Forward Error Correction (FEC) is enabled a network operator can choose to monitor an amount of corrected errors by using the pre-FEC mode, or the amount of errors which the FEC failed to correct (uncorrectable errors) by using the post-FEC mode, when FEC is used then every error detected by the PHY will be monitored.

When link is disabled the system will keep it in shutdown state until the port is explicitly enabled (Explicitly running “shutdown” and then “no shutdown” commands for that port).

Effective-BER is the BER that the MACLRH/Application layer observe. Errors monitored by the Effective-BER may directly result in a packet drop. For links with no error correction, the Effective BER is the BER received by port, and it is monitored based on the received Phy symbols. For links with FEC, the Effective BER represents the rate of errors that the FEC decoder did not manage to correct and were passed to the MACLRH layer. The Effective BER for FEC links is monitored using the FEC decoder uncorrectable codewords data.

  1. Enable signal degradation monitoring. Run:

    Copy
    Copied!
                

    switch (config) # interfaces ib 1/3 signal-degrade

    If not indicated, the interface is disabled in case of signal degradation.

  2. (Optional) To prevent the interface from shutting down in case of signal degradation, run:

    Copy
    Copied!
                

    switch (config) # interfaces ib 1/3 signal-degrade no-shutdown

    1. (Optional) Enable SNMP notifications on signal degradation events. Run:

      Copy
      Copied!
                  

      switch (config) # snmp notify event health-module-status

      Please refer to “Configuring SNMP Notifications (Traps or Informs)” for a general explanation on how to enable SNMP notifications for specific events.

  3. (Optional) Enable email notifications on signal degradation events. Run:

    Copy
    Copied!
                

    switch (config) # email notify event health-module-status

    Signal degradation snmp event comes only when there is an alarm alert of BER limit cross that is being sent only once. There is no SNMP alarm in case of cross down back to normal threshold, nor in the second time in a row the BER is crossed above again. In order to get another alarm on BER limit cross, it is needed to shutdown the interface and enable it again.

    Please refer to “Email Notifications” for a general explanation on how to enable email notifications for specific events.

signal-degrade

signal-degrade [no-shutdown]
no signal-degrade [no-shutdown]

Enables signal degradation operation per interface.
The no form of the command disables signal degradation operation per interface.

Syntax Description

no-shutdown

Does not shutdown an affected interface

Default

Disabled

Configuration Mode

config interface ib

History

3.6.6102

Example

switch (config interface ib 1/1) # signal-degrade

Related Commands

show interfaces ib signal-degrade

Notes

show interfaces ib signal-degrade

show interfaces ib [<slot>/<port>] signal-degrade

Displays signal degradation information.

Syntax Description

N/A

Default

N/A

Configuration Mode

config

History

3.6.6102

Example

switch (config) # show interfaces ib signal-degrade 
------------------------------------------------------------------------------------------
Interface Physical port state Monitoring Action FEC type
------------------------------------------------------------------------------------------
IB1/1 LinkUp Disabled Shutdown no-fec/post-fec
IB1/2 LinkUp Enabled Shutdown no-fec/post-fec
IB1/3 Polling Disabled Shutdown no-fec/post-fec
IB1/4 Polling Disabled Shutdown no-fec/post-fec
IB1/5 Polling Disabled Shutdown no-fec/post-fec
IB1/6 Polling Disabled Shutdown no-fec/post-fec
IB1/7 Polling Disabled Shutdown no-fec/post-fec
...

Related Commands

Notes

This command is relevant only for 1U switch systems

© Copyright 2023, NVIDIA. Last updated on Sep 8, 2023.